
from __future__ import absolute_import , division , print_function , unicode_literals

3 # TensorFlow and tf. keras
import tensorflow as tf
from tensorflow import keras
from random import seed
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
import pandas as pd
from sklearn.utils import shuffle
from tensorflow.keras.optimizers import Adagrad, Adam, RMSprop, SGD
from sklearn.model_selection import train_test_split
from keras.callbacks import ModelCheckpoint
from keras.callbacks import EarlyStopping
from keras.models import load_model
from keras.layers import BatchNormalization
from tensorflow.keras.utils import to_categorical
from sklearn.utils import shuffle
from tensorflow.keras import regularizers
from sklearn.metrics import accuracy_score
# create dataset


df=pd.DataFrame()
for i in range(1,36):
  globals()[f"df_{i}"] = pd.read_csv(f"heart_train_{i}.txt", sep=" ",header=None)
  df = pd.concat([df, globals()[f"df_{i}"]])
df


df=df.rename(columns={0:"age",1:"sex",2:"chest_pain",3:"blood_pressure",4:"cholestoral",5:"blood_sugar",6:"electrocardiographic",7:"max_heart_rate",8:"exercise",9:"oldpeak",10:"peak_exer",11:"nb_vessel",12:"defect",13:"disease"})
df


df.columns

Index(['age', 'sex', 'chest_pain', 'blood_pressure', 'cholestoral',
       'blood_sugar', 'electrocardiographic', 'max_heart_rate', 'exercise',
       'oldpeak', 'peak_exer', 'nb_vessel', 'defect', 'disease'],
      dtype='object')


df.to_csv("heart_train.txt",sep=" ", index=False)
#we save the dataset


df= pd.read_csv(f"heart_train.txt", sep=" ")
df


validation=pd.DataFrame()
for i in range(36,41):
  globals()[f"df_{i}"] = pd.read_csv(f"heart_train_{i}.txt", sep=" ",header=None)
  validation = pd.concat([validation, globals()[f"df_{i}"]])
validation


validation=validation.rename(columns={0:"age",1:"sex",2:"chest_pain",3:"blood_pressure",4:"cholestoral",5:"blood_sugar",6:"electrocardiographic",7:"max_heart_rate",8:"exercise",9:"oldpeak",10:"peak_exer",11:"nb_vessel",12:"defect",13:"disease"})
validation


validation.to_csv("heart_validation.txt",sep=" ", index=False)
#we save the dataset


small_df = pd.read_csv(f"heart_train_33.txt", sep=" ",header=None)
small_df=small_df.rename(columns={0:"age",1:"sex",2:"chest_pain",3:"blood_pressure",4:"cholestoral",5:"blood_sugar",6:"electrocardiographic",7:"max_heart_rate",8:"exercise",9:"oldpeak",10:"peak_exer",11:"nb_vessel",12:"defect",13:"disease"})
small_df


small_validation = pd.read_csv(f"heart_test.txt", sep=" ",header=None)
small_validation


small_validation=small_validation.rename(columns={0:"age",1:"sex",2:"chest_pain",3:"blood_pressure",4:"cholestoral",5:"blood_sugar",6:"electrocardiographic",7:"max_heart_rate",8:"exercise",9:"oldpeak",10:"peak_exer",11:"nb_vessel",12:"defect",13:"disease"})


df.describe()


sns.heatmap(df.corr(), cmap = 'coolwarm')

<Axes: >


sns.pairplot(df,hue='disease')

<seaborn.axisgrid.PairGrid at 0x7b65c00ff160>


df['disease'] = df['disease'].replace({1:0,2:1})
X = df.drop(columns="disease")
y=df["disease"]
y=y.values
y=to_categorical(y)
y.shape

(7560, 2)


X_train, X_test, Y_train, Y_test = train_test_split(X, y, test_size=0.3, random_state=1, stratify=y)


X_train[:5]


Y_train[:5]

array([[1., 0.],
       [1., 0.],
       [0., 1.],
       [1., 0.],
       [0., 1.]], dtype=float32)


print("X training set size:", len(X_train))
print("Y training set size:",len(Y_train))
print("X testing set size:",len(X_test))
print("Y testing set size:",len(Y_test))

X training set size: 5292
Y training set size: 5292
X testing set size: 2268
Y testing set size: 2268


mean = X_train.mean(axis=0)
std = X_train.std(axis=0)
X_train -= mean
X_train /= std
X_test -= mean
X_test /= std


small_df['disease'] = small_df['disease'].replace({1:0,2:1})
X = small_df.drop(columns="disease")
y=small_df["disease"]
y=y.values
y=to_categorical(y)
print("y shape : ",y.shape)
X_train, X_test, Y_train, Y_test = train_test_split(X, y, test_size=0.2, random_state=1, stratify=y)
#the dataset is really small so we reduce the test set to 20% of the dataset
print("X training set size:", len(X_train))
print("Y training set size:",len(Y_train))
print("X testing set size:",len(X_test))
print("Y testing set size:",len(Y_test))
mean = X_train.mean(axis=0)
std = X_train.std(axis=0)
X_train -= mean
X_train /= std
X_test -= mean
X_test /= std

y shape :  (216, 2)
X training set size: 172
Y training set size: 172
X testing set size: 44
Y testing set size: 44


def smooth_curve(points, factor=0.9):
  """function to plot the learning curves"""
  smoothed_points = []
  for point in points:
    if smoothed_points:
      previous = smoothed_points[-1]
      smoothed_points.append(previous * factor + point * (1 - factor))
    else:
      smoothed_points.append(point)
  return smoothed_points


def plot_history(history):
  """function to plot the learning curves of the model"""
  # plot learning curves
  plt.plot(history.history['accuracy'], label='train')
  plt.plot(history.history['val_accuracy'], label='test')
  plt.title('lrate='+str(lrate), pad=-50)
  plt.legend()
  plt.show()
  ####################################
  plt.plot(smooth_curve(history.history['accuracy']), label='train')
  plt.plot(smooth_curve(history.history['val_accuracy']), label='test')
  plt.title('lrate='+str(lrate), pad=-50)
  plt.legend()
  plt.show()
  #####################
  plt.plot(smooth_curve(history.history['loss']), label='train')
  plt.plot(smooth_curve(history.history['val_loss']), label='test')
  plt.title('lrate='+str(lrate), pad=-50)
  plt.legend()
  plt.show()


def creation_model0(X_train,Y_train,X_test,Y_test,earlystop,opti,lrate,l2reg):
  """function to select the optimization method of a model"""
  model = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform',kernel_regularizer=keras.regularizers.l2(l2reg)),
    #13 neurons because it's the number of columns/inputs
    keras.layers.Dense(2, activation='sigmoid')
    #2 neurons because we want a classification with two labels
])
  es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=earlystop)
  mc = ModelCheckpoint('best_model.h5', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
  if opti=="adam":
    opt=Adam(learning_rate= lrate)
  elif opti=="adagrad":
    opt=Adagrad(learning_rate= lrate)
  elif opti=="SGD":
    opt=SGD(learning_rate=0.1, momentum=0.9)
  else :
    opt= RMSprop(learning_rate= lrate, momentum=0.9)
  model.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])
  history = model.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0, callbacks=[es, mc])

  # plot learning curves
  plt.plot(history.history['accuracy'], label='train')
  plt.plot(history.history['val_accuracy'], label='test')
  plt.title('lrate='+str(lrate), pad=-50)
  plt.legend()
  plt.show()
  ####################################
  plt.plot(smooth_curve(history.history['accuracy']), label='train')
  plt.plot(smooth_curve(history.history['val_accuracy']), label='test')
  plt.title('lrate='+str(lrate), pad=-50)
  plt.legend()
  plt.show()
  #####################
  plt.plot(smooth_curve(history.history['loss']), label='train')
  plt.plot(smooth_curve(history.history['val_loss']), label='test')
  plt.title('lrate='+str(lrate), pad=-50)
  plt.legend()
  plt.show()


mod0 = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform'),
    #13 neurons because it's the number of columns/inputs
    keras.layers.Dense(2, activation='sigmoid')
    #2 neurons because we want a classification with two labels
])
es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=20)
mc = ModelCheckpoint('best_model.h0', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
lrate = 0.005
mod0.compile(loss='binary_crossentropy', optimizer=SGD(learning_rate= lrate), metrics=['accuracy'])
history = mod0.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0,callbacks=[es, mc])
saved_model = load_model('best_model.h0')

Epoch 1: val_accuracy improved from -inf to 0.63183, saving model to best_model.h0

Epoch 2: val_accuracy improved from 0.63183 to 0.70503, saving model to best_model.h0

Epoch 3: val_accuracy improved from 0.70503 to 0.75970, saving model to best_model.h0

Epoch 4: val_accuracy improved from 0.75970 to 0.79982, saving model to best_model.h0

Epoch 5: val_accuracy improved from 0.79982 to 0.80511, saving model to best_model.h0

Epoch 6: val_accuracy did not improve from 0.80511

Epoch 7: val_accuracy improved from 0.80511 to 0.82275, saving model to best_model.h0

Epoch 8: val_accuracy did not improve from 0.82275

Epoch 9: val_accuracy improved from 0.82275 to 0.82363, saving model to best_model.h0

Epoch 10: val_accuracy improved from 0.82363 to 0.84259, saving model to best_model.h0

Epoch 11: val_accuracy did not improve from 0.84259

Epoch 12: val_accuracy improved from 0.84259 to 0.84303, saving model to best_model.h0

Epoch 13: val_accuracy improved from 0.84303 to 0.84965, saving model to best_model.h0

Epoch 14: val_accuracy did not improve from 0.84965

Epoch 15: val_accuracy did not improve from 0.84965

Epoch 16: val_accuracy did not improve from 0.84965

Epoch 17: val_accuracy did not improve from 0.84965

Epoch 18: val_accuracy did not improve from 0.84965

Epoch 19: val_accuracy did not improve from 0.84965

Epoch 20: val_accuracy did not improve from 0.84965

Epoch 21: val_accuracy did not improve from 0.84965

Epoch 22: val_accuracy did not improve from 0.84965

Epoch 23: val_accuracy did not improve from 0.84965

Epoch 24: val_accuracy did not improve from 0.84965

Epoch 25: val_accuracy did not improve from 0.84965

Epoch 26: val_accuracy did not improve from 0.84965

Epoch 27: val_accuracy did not improve from 0.84965

Epoch 28: val_accuracy did not improve from 0.84965

Epoch 29: val_accuracy did not improve from 0.84965

Epoch 30: val_accuracy did not improve from 0.84965

Epoch 31: val_accuracy did not improve from 0.84965

Epoch 32: val_accuracy did not improve from 0.84965

Epoch 33: val_accuracy did not improve from 0.84965


mod0.summary()

Model: "sequential_2"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense_4 (Dense)             (None, 13)                182       
                                                                 
 dense_5 (Dense)             (None, 2)                 28        
                                                                 
=================================================================
Total params: 210 (840.00 Byte)
Trainable params: 210 (840.00 Byte)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________


plot_history(history)


mod1 = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    keras.layers.Dense(7, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    #mean(13+2) = 7.5 ~ 7 or 8 for the number of neurons in the hidden layer
    keras.layers.Dense(2, activation='sigmoid')
])
es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=35)
mc = ModelCheckpoint('best_model.h5', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
lrate = 0.005
mod1.compile(loss='binary_crossentropy', optimizer=SGD(learning_rate= lrate), metrics=['accuracy'])
history1 = mod1.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0,callbacks=[es, mc])

Epoch 1: val_accuracy improved from -inf to 0.57231, saving model to best_model.h5

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 2: val_accuracy improved from 0.57231 to 0.71252, saving model to best_model.h5

Epoch 3: val_accuracy improved from 0.71252 to 0.75132, saving model to best_model.h5

Epoch 4: val_accuracy improved from 0.75132 to 0.78483, saving model to best_model.h5

Epoch 5: val_accuracy improved from 0.78483 to 0.80511, saving model to best_model.h5

Epoch 6: val_accuracy improved from 0.80511 to 0.82496, saving model to best_model.h5

Epoch 7: val_accuracy improved from 0.82496 to 0.83113, saving model to best_model.h5

Epoch 8: val_accuracy did not improve from 0.83113

Epoch 9: val_accuracy did not improve from 0.83113

Epoch 10: val_accuracy improved from 0.83113 to 0.84171, saving model to best_model.h5

Epoch 11: val_accuracy improved from 0.84171 to 0.85009, saving model to best_model.h5

Epoch 12: val_accuracy improved from 0.85009 to 0.85582, saving model to best_model.h5

Epoch 13: val_accuracy did not improve from 0.85582

Epoch 14: val_accuracy improved from 0.85582 to 0.85979, saving model to best_model.h5

Epoch 15: val_accuracy did not improve from 0.85979

Epoch 16: val_accuracy improved from 0.85979 to 0.86464, saving model to best_model.h5

Epoch 17: val_accuracy did not improve from 0.86464

Epoch 18: val_accuracy improved from 0.86464 to 0.87522, saving model to best_model.h5

Epoch 19: val_accuracy improved from 0.87522 to 0.88492, saving model to best_model.h5

Epoch 20: val_accuracy improved from 0.88492 to 0.88889, saving model to best_model.h5

Epoch 21: val_accuracy improved from 0.88889 to 0.89242, saving model to best_model.h5

Epoch 22: val_accuracy did not improve from 0.89242

Epoch 23: val_accuracy improved from 0.89242 to 0.89771, saving model to best_model.h5

Epoch 24: val_accuracy did not improve from 0.89771

Epoch 25: val_accuracy did not improve from 0.89771

Epoch 26: val_accuracy improved from 0.89771 to 0.90123, saving model to best_model.h5

Epoch 27: val_accuracy improved from 0.90123 to 0.90653, saving model to best_model.h5

Epoch 28: val_accuracy improved from 0.90653 to 0.91446, saving model to best_model.h5

Epoch 29: val_accuracy improved from 0.91446 to 0.92240, saving model to best_model.h5

Epoch 30: val_accuracy improved from 0.92240 to 0.92593, saving model to best_model.h5

Epoch 31: val_accuracy improved from 0.92593 to 0.92945, saving model to best_model.h5

Epoch 32: val_accuracy did not improve from 0.92945

Epoch 33: val_accuracy did not improve from 0.92945

Epoch 34: val_accuracy did not improve from 0.92945

Epoch 35: val_accuracy improved from 0.92945 to 0.93430, saving model to best_model.h5

Epoch 36: val_accuracy did not improve from 0.93430

Epoch 37: val_accuracy improved from 0.93430 to 0.94444, saving model to best_model.h5

Epoch 38: val_accuracy did not improve from 0.94444

Epoch 39: val_accuracy did not improve from 0.94444

Epoch 40: val_accuracy did not improve from 0.94444

Epoch 41: val_accuracy improved from 0.94444 to 0.95018, saving model to best_model.h5

Epoch 42: val_accuracy improved from 0.95018 to 0.95194, saving model to best_model.h5

Epoch 43: val_accuracy improved from 0.95194 to 0.95767, saving model to best_model.h5

Epoch 44: val_accuracy did not improve from 0.95767

Epoch 45: val_accuracy did not improve from 0.95767

Epoch 46: val_accuracy did not improve from 0.95767

Epoch 47: val_accuracy did not improve from 0.95767

Epoch 48: val_accuracy improved from 0.95767 to 0.96296, saving model to best_model.h5

Epoch 49: val_accuracy improved from 0.96296 to 0.96384, saving model to best_model.h5

Epoch 50: val_accuracy did not improve from 0.96384

Epoch 51: val_accuracy improved from 0.96384 to 0.97840, saving model to best_model.h5

Epoch 52: val_accuracy did not improve from 0.97840

Epoch 53: val_accuracy did not improve from 0.97840

Epoch 54: val_accuracy did not improve from 0.97840

Epoch 55: val_accuracy did not improve from 0.97840

Epoch 56: val_accuracy improved from 0.97840 to 0.98457, saving model to best_model.h5

Epoch 57: val_accuracy did not improve from 0.98457

Epoch 58: val_accuracy did not improve from 0.98457

Epoch 59: val_accuracy did not improve from 0.98457

Epoch 60: val_accuracy did not improve from 0.98457

Epoch 61: val_accuracy did not improve from 0.98457

Epoch 62: val_accuracy did not improve from 0.98457

Epoch 63: val_accuracy did not improve from 0.98457

Epoch 64: val_accuracy did not improve from 0.98457

Epoch 65: val_accuracy improved from 0.98457 to 0.98942, saving model to best_model.h5

Epoch 66: val_accuracy did not improve from 0.98942

Epoch 67: val_accuracy did not improve from 0.98942

Epoch 68: val_accuracy did not improve from 0.98942

Epoch 69: val_accuracy did not improve from 0.98942

Epoch 70: val_accuracy did not improve from 0.98942

Epoch 71: val_accuracy did not improve from 0.98942

Epoch 72: val_accuracy improved from 0.98942 to 0.99471, saving model to best_model.h5

Epoch 73: val_accuracy did not improve from 0.99471

Epoch 74: val_accuracy did not improve from 0.99471

Epoch 75: val_accuracy did not improve from 0.99471

Epoch 76: val_accuracy did not improve from 0.99471

Epoch 77: val_accuracy did not improve from 0.99471

Epoch 78: val_accuracy did not improve from 0.99471

Epoch 79: val_accuracy did not improve from 0.99471

Epoch 80: val_accuracy did not improve from 0.99471

Epoch 81: val_accuracy did not improve from 0.99471

Epoch 82: val_accuracy did not improve from 0.99471

Epoch 83: val_accuracy did not improve from 0.99471

Epoch 84: val_accuracy did not improve from 0.99471

Epoch 85: val_accuracy did not improve from 0.99471

Epoch 86: val_accuracy did not improve from 0.99471

Epoch 87: val_accuracy did not improve from 0.99471

Epoch 88: val_accuracy did not improve from 0.99471

Epoch 89: val_accuracy did not improve from 0.99471

Epoch 90: val_accuracy did not improve from 0.99471

Epoch 91: val_accuracy did not improve from 0.99471

Epoch 92: val_accuracy did not improve from 0.99471

Epoch 93: val_accuracy did not improve from 0.99471

Epoch 94: val_accuracy did not improve from 0.99471

Epoch 95: val_accuracy did not improve from 0.99471

Epoch 96: val_accuracy did not improve from 0.99471

Epoch 97: val_accuracy did not improve from 0.99471

Epoch 98: val_accuracy did not improve from 0.99471

Epoch 99: val_accuracy did not improve from 0.99471

Epoch 100: val_accuracy did not improve from 0.99471

Epoch 101: val_accuracy did not improve from 0.99471

Epoch 102: val_accuracy did not improve from 0.99471

Epoch 103: val_accuracy did not improve from 0.99471

Epoch 104: val_accuracy did not improve from 0.99471

Epoch 105: val_accuracy did not improve from 0.99471

Epoch 106: val_accuracy did not improve from 0.99471

Epoch 107: val_accuracy did not improve from 0.99471


mod1.summary()

Model: "sequential_9"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense_24 (Dense)            (None, 13)                182       
                                                                 
 batch_normalization_12 (Ba  (None, 13)                52        
 tchNormalization)                                               
                                                                 
 dense_25 (Dense)            (None, 7)                 98        
                                                                 
 batch_normalization_13 (Ba  (None, 7)                 28        
 tchNormalization)                                               
                                                                 
 dense_26 (Dense)            (None, 2)                 16        
                                                                 
=================================================================
Total params: 376 (1.47 KB)
Trainable params: 336 (1.31 KB)
Non-trainable params: 40 (160.00 Byte)
_________________________________________________________________


plot_history(history1)


mod2 = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    keras.layers.Dense(8, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    #mean(13+2) = 7.5 ~ 7 or 8 for the number of neurons in the hidden layer
    keras.layers.Dense(2, activation='sigmoid')
])
es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=20)
mc = ModelCheckpoint('best_model.h5', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
lrate = 0.005
mod2.compile(loss='binary_crossentropy', optimizer=SGD(learning_rate= lrate), metrics=['accuracy'])
history2 = mod2.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0,callbacks=[es, mc])

Epoch 1: val_accuracy improved from -inf to 0.73677, saving model to best_model.h5

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 2: val_accuracy improved from 0.73677 to 0.73765, saving model to best_model.h5

Epoch 3: val_accuracy improved from 0.73765 to 0.76455, saving model to best_model.h5

Epoch 4: val_accuracy improved from 0.76455 to 0.79189, saving model to best_model.h5

Epoch 5: val_accuracy improved from 0.79189 to 0.80291, saving model to best_model.h5

Epoch 6: val_accuracy did not improve from 0.80291

Epoch 7: val_accuracy improved from 0.80291 to 0.80423, saving model to best_model.h5

Epoch 8: val_accuracy improved from 0.80423 to 0.82099, saving model to best_model.h5

Epoch 9: val_accuracy did not improve from 0.82099

Epoch 10: val_accuracy improved from 0.82099 to 0.82407, saving model to best_model.h5

Epoch 11: val_accuracy improved from 0.82407 to 0.82937, saving model to best_model.h5

Epoch 12: val_accuracy improved from 0.82937 to 0.83466, saving model to best_model.h5

Epoch 13: val_accuracy improved from 0.83466 to 0.83862, saving model to best_model.h5

Epoch 14: val_accuracy improved from 0.83862 to 0.84259, saving model to best_model.h5

Epoch 15: val_accuracy improved from 0.84259 to 0.84524, saving model to best_model.h5

Epoch 16: val_accuracy improved from 0.84524 to 0.86243, saving model to best_model.h5

Epoch 17: val_accuracy did not improve from 0.86243

Epoch 18: val_accuracy improved from 0.86243 to 0.86772, saving model to best_model.h5

Epoch 19: val_accuracy improved from 0.86772 to 0.87831, saving model to best_model.h5

Epoch 20: val_accuracy improved from 0.87831 to 0.88757, saving model to best_model.h5

Epoch 21: val_accuracy did not improve from 0.88757

Epoch 22: val_accuracy improved from 0.88757 to 0.89109, saving model to best_model.h5

Epoch 23: val_accuracy improved from 0.89109 to 0.89374, saving model to best_model.h5

Epoch 24: val_accuracy did not improve from 0.89374

Epoch 25: val_accuracy improved from 0.89374 to 0.90873, saving model to best_model.h5

Epoch 26: val_accuracy improved from 0.90873 to 0.92240, saving model to best_model.h5

Epoch 27: val_accuracy did not improve from 0.92240

Epoch 28: val_accuracy did not improve from 0.92240

Epoch 29: val_accuracy did not improve from 0.92240

Epoch 30: val_accuracy did not improve from 0.92240

Epoch 31: val_accuracy did not improve from 0.92240

Epoch 32: val_accuracy improved from 0.92240 to 0.92593, saving model to best_model.h5

Epoch 33: val_accuracy improved from 0.92593 to 0.92945, saving model to best_model.h5

Epoch 34: val_accuracy improved from 0.92945 to 0.93563, saving model to best_model.h5

Epoch 35: val_accuracy did not improve from 0.93563

Epoch 36: val_accuracy did not improve from 0.93563

Epoch 37: val_accuracy did not improve from 0.93563

Epoch 38: val_accuracy did not improve from 0.93563

Epoch 39: val_accuracy improved from 0.93563 to 0.94577, saving model to best_model.h5

Epoch 40: val_accuracy improved from 0.94577 to 0.94797, saving model to best_model.h5

Epoch 41: val_accuracy improved from 0.94797 to 0.95062, saving model to best_model.h5

Epoch 42: val_accuracy did not improve from 0.95062

Epoch 43: val_accuracy did not improve from 0.95062

Epoch 44: val_accuracy did not improve from 0.95062

Epoch 45: val_accuracy did not improve from 0.95062

Epoch 46: val_accuracy improved from 0.95062 to 0.95767, saving model to best_model.h5

Epoch 47: val_accuracy did not improve from 0.95767

Epoch 48: val_accuracy did not improve from 0.95767

Epoch 49: val_accuracy did not improve from 0.95767

Epoch 50: val_accuracy did not improve from 0.95767

Epoch 51: val_accuracy did not improve from 0.95767

Epoch 52: val_accuracy improved from 0.95767 to 0.96384, saving model to best_model.h5

Epoch 53: val_accuracy did not improve from 0.96384

Epoch 54: val_accuracy did not improve from 0.96384

Epoch 55: val_accuracy did not improve from 0.96384

Epoch 56: val_accuracy did not improve from 0.96384

Epoch 57: val_accuracy did not improve from 0.96384

Epoch 58: val_accuracy did not improve from 0.96384

Epoch 59: val_accuracy did not improve from 0.96384

Epoch 60: val_accuracy improved from 0.96384 to 0.96649, saving model to best_model.h5

Epoch 61: val_accuracy did not improve from 0.96649

Epoch 62: val_accuracy did not improve from 0.96649

Epoch 63: val_accuracy did not improve from 0.96649

Epoch 64: val_accuracy did not improve from 0.96649

Epoch 65: val_accuracy did not improve from 0.96649

Epoch 66: val_accuracy did not improve from 0.96649

Epoch 67: val_accuracy improved from 0.96649 to 0.97840, saving model to best_model.h5

Epoch 68: val_accuracy did not improve from 0.97840

Epoch 69: val_accuracy did not improve from 0.97840

Epoch 70: val_accuracy did not improve from 0.97840

Epoch 71: val_accuracy did not improve from 0.97840

Epoch 72: val_accuracy did not improve from 0.97840

Epoch 73: val_accuracy did not improve from 0.97840

Epoch 74: val_accuracy did not improve from 0.97840

Epoch 75: val_accuracy did not improve from 0.97840

Epoch 76: val_accuracy improved from 0.97840 to 0.98192, saving model to best_model.h5

Epoch 77: val_accuracy did not improve from 0.98192

Epoch 78: val_accuracy did not improve from 0.98192

Epoch 79: val_accuracy did not improve from 0.98192

Epoch 80: val_accuracy did not improve from 0.98192

Epoch 81: val_accuracy did not improve from 0.98192

Epoch 82: val_accuracy did not improve from 0.98192

Epoch 83: val_accuracy did not improve from 0.98192

Epoch 84: val_accuracy did not improve from 0.98192

Epoch 85: val_accuracy did not improve from 0.98192

Epoch 86: val_accuracy improved from 0.98192 to 0.98457, saving model to best_model.h5

Epoch 87: val_accuracy did not improve from 0.98457

Epoch 88: val_accuracy did not improve from 0.98457

Epoch 89: val_accuracy did not improve from 0.98457

Epoch 90: val_accuracy did not improve from 0.98457

Epoch 91: val_accuracy did not improve from 0.98457

Epoch 92: val_accuracy did not improve from 0.98457

Epoch 93: val_accuracy did not improve from 0.98457

Epoch 94: val_accuracy did not improve from 0.98457

Epoch 95: val_accuracy did not improve from 0.98457

Epoch 96: val_accuracy did not improve from 0.98457

Epoch 97: val_accuracy did not improve from 0.98457

Epoch 98: val_accuracy did not improve from 0.98457

Epoch 99: val_accuracy did not improve from 0.98457

Epoch 100: val_accuracy did not improve from 0.98457

Epoch 101: val_accuracy did not improve from 0.98457

Epoch 102: val_accuracy did not improve from 0.98457

Epoch 103: val_accuracy did not improve from 0.98457

Epoch 104: val_accuracy did not improve from 0.98457

Epoch 105: val_accuracy did not improve from 0.98457

Epoch 106: val_accuracy improved from 0.98457 to 0.98986, saving model to best_model.h5

Epoch 107: val_accuracy improved from 0.98986 to 0.99515, saving model to best_model.h5

Epoch 108: val_accuracy did not improve from 0.99515

Epoch 109: val_accuracy did not improve from 0.99515

Epoch 110: val_accuracy did not improve from 0.99515

Epoch 111: val_accuracy did not improve from 0.99515

Epoch 112: val_accuracy did not improve from 0.99515

Epoch 113: val_accuracy did not improve from 0.99515

Epoch 114: val_accuracy did not improve from 0.99515

Epoch 115: val_accuracy did not improve from 0.99515

Epoch 116: val_accuracy did not improve from 0.99515

Epoch 117: val_accuracy did not improve from 0.99515

Epoch 118: val_accuracy did not improve from 0.99515

Epoch 119: val_accuracy did not improve from 0.99515

Epoch 120: val_accuracy did not improve from 0.99515

Epoch 121: val_accuracy did not improve from 0.99515

Epoch 122: val_accuracy did not improve from 0.99515

Epoch 123: val_accuracy did not improve from 0.99515

Epoch 124: val_accuracy did not improve from 0.99515

Epoch 125: val_accuracy did not improve from 0.99515

Epoch 126: val_accuracy did not improve from 0.99515

Epoch 127: val_accuracy did not improve from 0.99515


mod2.summary()

Model: "sequential_10"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense_27 (Dense)            (None, 13)                182       
                                                                 
 batch_normalization_14 (Ba  (None, 13)                52        
 tchNormalization)                                               
                                                                 
 dense_28 (Dense)            (None, 8)                 112       
                                                                 
 batch_normalization_15 (Ba  (None, 8)                 32        
 tchNormalization)                                               
                                                                 
 dense_29 (Dense)            (None, 2)                 18        
                                                                 
=================================================================
Total params: 396 (1.55 KB)
Trainable params: 354 (1.38 KB)
Non-trainable params: 42 (168.00 Byte)
_________________________________________________________________


plot_history(history2)


X = validation.drop(columns="disease")
y=validation["disease"]

y.shape

(1080,)


X-= mean
X /= std


predictions = mod2.predict(X)

34/34 [==============================] - 0s 2ms/step


predictions[:5]

array([[8.4876359e-01, 1.2433447e-01],
       [3.5142712e-02, 9.6764618e-01],
       [1.7565981e-04, 9.9975723e-01],
       [9.9389571e-01, 4.6091015e-03],
       [9.8444611e-01, 1.3892172e-02]], dtype=float32)


result =[]
for i in range(len(predictions)):
  if predictions[i][0]>predictions[i][1]:
    result.append(1)#the real class
  else :
    result.append(2)#the real class


score = accuracy_score(validation["disease"], result)
print("accuracy : "+str(score))

accuracy : 0.9953703703703703


#writing results for exam evaluation
r= open("results_big_dataset.txt", "w+")
r.write(str(result))
r.close()


mod0 = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform'),
    #13 neurons because it's the number of columns/inputs
    keras.layers.Dense(2, activation='sigmoid')
    #2 neurons because we want a classification with two labels
])
es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=20)
mc = ModelCheckpoint('best_model.h0', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
lrate = 0.005
mod0.compile(loss='binary_crossentropy', optimizer=SGD(learning_rate= lrate), metrics=['accuracy'])
history = mod0.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0,callbacks=[es, mc])

Epoch 1: val_accuracy improved from -inf to 0.43182, saving model to best_model.h0

Epoch 2: val_accuracy did not improve from 0.43182

Epoch 3: val_accuracy did not improve from 0.43182

Epoch 4: val_accuracy improved from 0.43182 to 0.45455, saving model to best_model.h0

Epoch 5: val_accuracy did not improve from 0.45455

Epoch 6: val_accuracy improved from 0.45455 to 0.47727, saving model to best_model.h0

Epoch 7: val_accuracy did not improve from 0.47727

Epoch 8: val_accuracy did not improve from 0.47727

Epoch 9: val_accuracy did not improve from 0.47727

Epoch 10: val_accuracy did not improve from 0.47727

Epoch 11: val_accuracy did not improve from 0.47727

Epoch 12: val_accuracy did not improve from 0.47727

Epoch 13: val_accuracy improved from 0.47727 to 0.50000, saving model to best_model.h0

Epoch 14: val_accuracy did not improve from 0.50000

Epoch 15: val_accuracy did not improve from 0.50000

Epoch 16: val_accuracy did not improve from 0.50000

Epoch 17: val_accuracy did not improve from 0.50000

Epoch 18: val_accuracy did not improve from 0.50000

Epoch 19: val_accuracy did not improve from 0.50000

Epoch 20: val_accuracy did not improve from 0.50000

Epoch 21: val_accuracy did not improve from 0.50000

Epoch 22: val_accuracy improved from 0.50000 to 0.52273, saving model to best_model.h0

Epoch 23: val_accuracy improved from 0.52273 to 0.54545, saving model to best_model.h0

Epoch 24: val_accuracy did not improve from 0.54545

Epoch 25: val_accuracy improved from 0.54545 to 0.56818, saving model to best_model.h0

Epoch 26: val_accuracy did not improve from 0.56818

Epoch 27: val_accuracy improved from 0.56818 to 0.59091, saving model to best_model.h0

Epoch 28: val_accuracy did not improve from 0.59091

Epoch 29: val_accuracy did not improve from 0.59091

Epoch 30: val_accuracy did not improve from 0.59091

Epoch 31: val_accuracy improved from 0.59091 to 0.61364, saving model to best_model.h0

Epoch 32: val_accuracy did not improve from 0.61364

Epoch 33: val_accuracy improved from 0.61364 to 0.63636, saving model to best_model.h0

Epoch 34: val_accuracy did not improve from 0.63636

Epoch 35: val_accuracy did not improve from 0.63636

Epoch 36: val_accuracy did not improve from 0.63636

Epoch 37: val_accuracy did not improve from 0.63636

Epoch 38: val_accuracy improved from 0.63636 to 0.65909, saving model to best_model.h0

Epoch 39: val_accuracy did not improve from 0.65909

Epoch 40: val_accuracy did not improve from 0.65909

Epoch 41: val_accuracy did not improve from 0.65909

Epoch 42: val_accuracy did not improve from 0.65909

Epoch 43: val_accuracy did not improve from 0.65909

Epoch 44: val_accuracy did not improve from 0.65909

Epoch 45: val_accuracy did not improve from 0.65909

Epoch 46: val_accuracy did not improve from 0.65909

Epoch 47: val_accuracy did not improve from 0.65909

Epoch 48: val_accuracy did not improve from 0.65909

Epoch 49: val_accuracy did not improve from 0.65909

Epoch 50: val_accuracy did not improve from 0.65909

Epoch 51: val_accuracy improved from 0.65909 to 0.68182, saving model to best_model.h0

Epoch 52: val_accuracy did not improve from 0.68182

Epoch 53: val_accuracy did not improve from 0.68182

Epoch 54: val_accuracy did not improve from 0.68182

Epoch 55: val_accuracy did not improve from 0.68182

Epoch 56: val_accuracy did not improve from 0.68182

Epoch 57: val_accuracy did not improve from 0.68182

Epoch 58: val_accuracy did not improve from 0.68182

Epoch 59: val_accuracy did not improve from 0.68182

Epoch 60: val_accuracy did not improve from 0.68182

Epoch 61: val_accuracy did not improve from 0.68182

Epoch 62: val_accuracy did not improve from 0.68182

Epoch 63: val_accuracy did not improve from 0.68182

Epoch 64: val_accuracy did not improve from 0.68182

Epoch 65: val_accuracy did not improve from 0.68182

Epoch 66: val_accuracy did not improve from 0.68182

Epoch 67: val_accuracy did not improve from 0.68182

Epoch 68: val_accuracy did not improve from 0.68182

Epoch 69: val_accuracy did not improve from 0.68182

Epoch 70: val_accuracy improved from 0.68182 to 0.70455, saving model to best_model.h0

Epoch 71: val_accuracy did not improve from 0.70455

Epoch 72: val_accuracy did not improve from 0.70455

Epoch 73: val_accuracy improved from 0.70455 to 0.72727, saving model to best_model.h0

Epoch 74: val_accuracy did not improve from 0.72727

Epoch 75: val_accuracy did not improve from 0.72727

Epoch 76: val_accuracy did not improve from 0.72727

Epoch 77: val_accuracy improved from 0.72727 to 0.77273, saving model to best_model.h0

Epoch 78: val_accuracy did not improve from 0.77273

Epoch 79: val_accuracy did not improve from 0.77273

Epoch 80: val_accuracy did not improve from 0.77273

Epoch 81: val_accuracy did not improve from 0.77273

Epoch 82: val_accuracy did not improve from 0.77273

Epoch 83: val_accuracy did not improve from 0.77273

Epoch 84: val_accuracy did not improve from 0.77273

Epoch 85: val_accuracy did not improve from 0.77273

Epoch 86: val_accuracy improved from 0.77273 to 0.79545, saving model to best_model.h0

Epoch 87: val_accuracy did not improve from 0.79545

Epoch 88: val_accuracy did not improve from 0.79545

Epoch 89: val_accuracy did not improve from 0.79545

Epoch 90: val_accuracy did not improve from 0.79545

Epoch 91: val_accuracy did not improve from 0.79545

Epoch 92: val_accuracy did not improve from 0.79545

Epoch 93: val_accuracy did not improve from 0.79545

Epoch 94: val_accuracy did not improve from 0.79545

Epoch 95: val_accuracy did not improve from 0.79545

Epoch 96: val_accuracy did not improve from 0.79545

Epoch 97: val_accuracy did not improve from 0.79545

Epoch 98: val_accuracy did not improve from 0.79545

Epoch 99: val_accuracy did not improve from 0.79545

Epoch 100: val_accuracy did not improve from 0.79545

Epoch 101: val_accuracy did not improve from 0.79545

Epoch 102: val_accuracy did not improve from 0.79545

Epoch 103: val_accuracy did not improve from 0.79545

Epoch 104: val_accuracy did not improve from 0.79545

Epoch 105: val_accuracy did not improve from 0.79545

Epoch 106: val_accuracy did not improve from 0.79545


mod0.summary()

Model: "sequential_12"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense_32 (Dense)            (None, 13)                182       
                                                                 
 dense_33 (Dense)            (None, 2)                 28        
                                                                 
=================================================================
Total params: 210 (840.00 Byte)
Trainable params: 210 (840.00 Byte)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________


plot_history(history)


mod1 = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    keras.layers.Dense(7, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    #mean(13+2) = 7.5 ~ 7 or 8 for the number of neurons in the hidden layer
    keras.layers.Dense(2, activation='sigmoid')
])
es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=20)
mc = ModelCheckpoint('best_model.h5', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
lrate = 0.005
mod1.compile(loss='binary_crossentropy', optimizer=SGD(learning_rate= lrate), metrics=['accuracy'])
history1 = mod1.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0,callbacks=[es, mc])

Epoch 1: val_accuracy improved from -inf to 0.47727, saving model to best_model.h5

Epoch 2: val_accuracy did not improve from 0.47727

Epoch 3: val_accuracy improved from 0.47727 to 0.52273, saving model to best_model.h5

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 4: val_accuracy improved from 0.52273 to 0.54545, saving model to best_model.h5

Epoch 5: val_accuracy did not improve from 0.54545

Epoch 6: val_accuracy did not improve from 0.54545

Epoch 7: val_accuracy improved from 0.54545 to 0.56818, saving model to best_model.h5

Epoch 8: val_accuracy did not improve from 0.56818

Epoch 9: val_accuracy did not improve from 0.56818

Epoch 10: val_accuracy did not improve from 0.56818

Epoch 11: val_accuracy did not improve from 0.56818

Epoch 12: val_accuracy did not improve from 0.56818

Epoch 13: val_accuracy did not improve from 0.56818

Epoch 14: val_accuracy improved from 0.56818 to 0.59091, saving model to best_model.h5

Epoch 15: val_accuracy improved from 0.59091 to 0.61364, saving model to best_model.h5

Epoch 16: val_accuracy did not improve from 0.61364

Epoch 17: val_accuracy did not improve from 0.61364

Epoch 18: val_accuracy did not improve from 0.61364

Epoch 19: val_accuracy did not improve from 0.61364

Epoch 20: val_accuracy did not improve from 0.61364

Epoch 21: val_accuracy did not improve from 0.61364

Epoch 22: val_accuracy did not improve from 0.61364

Epoch 23: val_accuracy did not improve from 0.61364

Epoch 24: val_accuracy did not improve from 0.61364

Epoch 25: val_accuracy did not improve from 0.61364

Epoch 26: val_accuracy did not improve from 0.61364

Epoch 27: val_accuracy did not improve from 0.61364

Epoch 28: val_accuracy did not improve from 0.61364

Epoch 29: val_accuracy did not improve from 0.61364

Epoch 30: val_accuracy did not improve from 0.61364

Epoch 31: val_accuracy did not improve from 0.61364

Epoch 32: val_accuracy did not improve from 0.61364

Epoch 33: val_accuracy did not improve from 0.61364

Epoch 34: val_accuracy did not improve from 0.61364

Epoch 35: val_accuracy did not improve from 0.61364


mod1.summary()

Model: "sequential_14"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense_37 (Dense)            (None, 13)                182       
                                                                 
 batch_normalization_18 (Ba  (None, 13)                52        
 tchNormalization)                                               
                                                                 
 dense_38 (Dense)            (None, 7)                 98        
                                                                 
 batch_normalization_19 (Ba  (None, 7)                 28        
 tchNormalization)                                               
                                                                 
 dense_39 (Dense)            (None, 2)                 16        
                                                                 
=================================================================
Total params: 376 (1.47 KB)
Trainable params: 336 (1.31 KB)
Non-trainable params: 40 (160.00 Byte)
_________________________________________________________________


plot_history(history1)


mod2 = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    keras.layers.Dense(8, activation='relu', kernel_initializer='he_uniform'),BatchNormalization(),
    #mean(13+2) = 7.5 ~ 7 or 8 for the number of neurons in the hidden layer
    keras.layers.Dense(2, activation='sigmoid')
])
es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=20)
mc = ModelCheckpoint('best_model.h5', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
lrate = 0.005
mod2.compile(loss='binary_crossentropy', optimizer=SGD(learning_rate= lrate), metrics=['accuracy'])
history2 = mod2.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0,callbacks=[es, mc])

Epoch 1: val_accuracy improved from -inf to 0.70455, saving model to best_model.h5

Epoch 2: val_accuracy did not improve from 0.70455

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 3: val_accuracy did not improve from 0.70455

Epoch 4: val_accuracy did not improve from 0.70455

Epoch 5: val_accuracy did not improve from 0.70455

Epoch 6: val_accuracy did not improve from 0.70455

Epoch 7: val_accuracy did not improve from 0.70455

Epoch 8: val_accuracy did not improve from 0.70455

Epoch 9: val_accuracy did not improve from 0.70455

Epoch 10: val_accuracy improved from 0.70455 to 0.72727, saving model to best_model.h5

Epoch 11: val_accuracy did not improve from 0.72727

Epoch 12: val_accuracy did not improve from 0.72727

Epoch 13: val_accuracy did not improve from 0.72727

Epoch 14: val_accuracy did not improve from 0.72727

Epoch 15: val_accuracy did not improve from 0.72727

Epoch 16: val_accuracy did not improve from 0.72727

Epoch 17: val_accuracy did not improve from 0.72727

Epoch 18: val_accuracy did not improve from 0.72727

Epoch 19: val_accuracy did not improve from 0.72727

Epoch 20: val_accuracy did not improve from 0.72727

Epoch 21: val_accuracy improved from 0.72727 to 0.75000, saving model to best_model.h5

Epoch 22: val_accuracy did not improve from 0.75000

Epoch 23: val_accuracy did not improve from 0.75000

Epoch 24: val_accuracy improved from 0.75000 to 0.77273, saving model to best_model.h5

Epoch 25: val_accuracy did not improve from 0.77273

Epoch 26: val_accuracy did not improve from 0.77273

Epoch 27: val_accuracy did not improve from 0.77273

Epoch 28: val_accuracy did not improve from 0.77273

Epoch 29: val_accuracy did not improve from 0.77273

Epoch 30: val_accuracy did not improve from 0.77273

Epoch 31: val_accuracy did not improve from 0.77273

Epoch 32: val_accuracy improved from 0.77273 to 0.79545, saving model to best_model.h5

Epoch 33: val_accuracy did not improve from 0.79545

Epoch 34: val_accuracy did not improve from 0.79545

Epoch 35: val_accuracy did not improve from 0.79545

Epoch 36: val_accuracy improved from 0.79545 to 0.81818, saving model to best_model.h5

Epoch 37: val_accuracy did not improve from 0.81818

Epoch 38: val_accuracy did not improve from 0.81818

Epoch 39: val_accuracy did not improve from 0.81818

Epoch 40: val_accuracy did not improve from 0.81818

Epoch 41: val_accuracy did not improve from 0.81818

Epoch 42: val_accuracy did not improve from 0.81818

Epoch 43: val_accuracy improved from 0.81818 to 0.84091, saving model to best_model.h5

Epoch 44: val_accuracy did not improve from 0.84091

Epoch 45: val_accuracy did not improve from 0.84091

Epoch 46: val_accuracy did not improve from 0.84091

Epoch 47: val_accuracy did not improve from 0.84091

Epoch 48: val_accuracy did not improve from 0.84091

Epoch 49: val_accuracy did not improve from 0.84091

Epoch 50: val_accuracy did not improve from 0.84091

Epoch 51: val_accuracy did not improve from 0.84091

Epoch 52: val_accuracy did not improve from 0.84091

Epoch 53: val_accuracy did not improve from 0.84091

Epoch 54: val_accuracy did not improve from 0.84091

Epoch 55: val_accuracy did not improve from 0.84091

Epoch 56: val_accuracy did not improve from 0.84091

Epoch 57: val_accuracy did not improve from 0.84091

Epoch 58: val_accuracy did not improve from 0.84091

Epoch 59: val_accuracy improved from 0.84091 to 0.86364, saving model to best_model.h5

Epoch 60: val_accuracy did not improve from 0.86364

Epoch 61: val_accuracy did not improve from 0.86364

Epoch 62: val_accuracy did not improve from 0.86364

Epoch 63: val_accuracy did not improve from 0.86364

Epoch 64: val_accuracy did not improve from 0.86364

Epoch 65: val_accuracy did not improve from 0.86364

Epoch 66: val_accuracy did not improve from 0.86364

Epoch 67: val_accuracy did not improve from 0.86364

Epoch 68: val_accuracy did not improve from 0.86364

Epoch 69: val_accuracy did not improve from 0.86364

Epoch 70: val_accuracy did not improve from 0.86364

Epoch 71: val_accuracy did not improve from 0.86364

Epoch 72: val_accuracy did not improve from 0.86364

Epoch 73: val_accuracy did not improve from 0.86364

Epoch 74: val_accuracy did not improve from 0.86364

Epoch 75: val_accuracy did not improve from 0.86364

Epoch 76: val_accuracy did not improve from 0.86364

Epoch 77: val_accuracy did not improve from 0.86364

Epoch 78: val_accuracy did not improve from 0.86364

Epoch 79: val_accuracy did not improve from 0.86364


mod2.summary()

Model: "sequential_15"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense_40 (Dense)            (None, 13)                182       
                                                                 
 batch_normalization_20 (Ba  (None, 13)                52        
 tchNormalization)                                               
                                                                 
 dense_41 (Dense)            (None, 8)                 112       
                                                                 
 batch_normalization_21 (Ba  (None, 8)                 32        
 tchNormalization)                                               
                                                                 
 dense_42 (Dense)            (None, 2)                 18        
                                                                 
=================================================================
Total params: 396 (1.55 KB)
Trainable params: 354 (1.38 KB)
Non-trainable params: 42 (168.00 Byte)
_________________________________________________________________


plot_history(history2)


learning_rates = [1E-0, 1E-1, 1E-2, 1E-3,0.005,0.0025, 1E-4, 1E-5, 1E-6, 1E-7]
l2reg = [1,0.1,0.01,0.001,0.0001,0.0001]
classement =[]
ind=3
for k in range(len(l2reg)):
  for j,opti in enumerate(["adam","adagrad","SGD"]):
    for i in range(len(learning_rates)):
      print("\n\n#######################################################\n\n")
      print(f"the model mod{ind} use a learning rate = {i}, l2 regularization = {k} and the optimizer = {opti} :")
      try :
        globals()[f"mod{ind}"] = creation_model0(X_train,Y_train,X_test,Y_test,20,opti,learning_rates[i],l2reg[k])
        classement.append([globals()[f"mod{ind}"].evaluate(X_test,  Y_test, verbose=2)[1],f"mod{ind}"])

      except Exception as e:
        print(e)
      ind+=1


#######################################################


the model mod3 use a learning rate = 0, l2 regularization = 0 and the optimizer = adam :

Epoch 1: val_accuracy improved from -inf to 0.54545, saving model to best_model.h5

Epoch 2: val_accuracy did not improve from 0.54545

Epoch 3: val_accuracy did not improve from 0.54545

Epoch 4: val_accuracy did not improve from 0.54545

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 5: val_accuracy did not improve from 0.54545

Epoch 6: val_accuracy did not improve from 0.54545

Epoch 7: val_accuracy did not improve from 0.54545

Epoch 8: val_accuracy did not improve from 0.54545

Epoch 9: val_accuracy did not improve from 0.54545

Epoch 10: val_accuracy did not improve from 0.54545

Epoch 11: val_accuracy did not improve from 0.54545

Epoch 12: val_accuracy did not improve from 0.54545

Epoch 13: val_accuracy did not improve from 0.54545

Epoch 14: val_accuracy did not improve from 0.54545

Epoch 15: val_accuracy did not improve from 0.54545

Epoch 16: val_accuracy did not improve from 0.54545

Epoch 17: val_accuracy did not improve from 0.54545

Epoch 18: val_accuracy did not improve from 0.54545

Epoch 19: val_accuracy did not improve from 0.54545

Epoch 20: val_accuracy did not improve from 0.54545

Epoch 21: val_accuracy did not improve from 0.54545

2/2 - 0s - loss: 0.6909 - accuracy: 0.5455 - 24ms/epoch - 12ms/step


#######################################################


the model mod4 use a learning rate = 1, l2 regularization = 0 and the optimizer = adam :

Epoch 1: val_accuracy improved from -inf to 0.54545, saving model to best_model.h5

Epoch 2: val_accuracy did not improve from 0.54545

Epoch 3: val_accuracy improved from 0.54545 to 0.72727, saving model to best_model.h5

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 4: val_accuracy improved from 0.72727 to 0.88636, saving model to best_model.h5

Epoch 5: val_accuracy did not improve from 0.88636

Epoch 6: val_accuracy did not improve from 0.88636

Epoch 7: val_accuracy did not improve from 0.88636

Epoch 8: val_accuracy did not improve from 0.88636

Epoch 9: val_accuracy did not improve from 0.88636

Epoch 10: val_accuracy did not improve from 0.88636

Epoch 11: val_accuracy did not improve from 0.88636

Epoch 12: val_accuracy did not improve from 0.88636

Epoch 13: val_accuracy did not improve from 0.88636

Epoch 14: val_accuracy did not improve from 0.88636

Epoch 15: val_accuracy did not improve from 0.88636

Epoch 16: val_accuracy did not improve from 0.88636

Epoch 17: val_accuracy did not improve from 0.88636

Epoch 18: val_accuracy improved from 0.88636 to 0.90909, saving model to best_model.h5

Epoch 19: val_accuracy did not improve from 0.90909

Epoch 20: val_accuracy did not improve from 0.90909

Epoch 21: val_accuracy did not improve from 0.90909

Epoch 22: val_accuracy did not improve from 0.90909

Epoch 23: val_accuracy did not improve from 0.90909

Epoch 24: val_accuracy did not improve from 0.90909

Epoch 25: val_accuracy did not improve from 0.90909

Epoch 26: val_accuracy did not improve from 0.90909

Epoch 27: val_accuracy did not improve from 0.90909

Epoch 28: val_accuracy did not improve from 0.90909

Epoch 29: val_accuracy did not improve from 0.90909

Epoch 30: val_accuracy did not improve from 0.90909

Epoch 31: val_accuracy did not improve from 0.90909

Epoch 32: val_accuracy did not improve from 0.90909

Epoch 33: val_accuracy did not improve from 0.90909

Epoch 34: val_accuracy did not improve from 0.90909

Epoch 35: val_accuracy did not improve from 0.90909

Epoch 36: val_accuracy did not improve from 0.90909

Epoch 37: val_accuracy did not improve from 0.90909

Epoch 38: val_accuracy did not improve from 0.90909

2/2 - 0s - loss: 0.5070 - accuracy: 0.8636 - 88ms/epoch - 44ms/step


#######################################################


the model mod5 use a learning rate = 2, l2 regularization = 0 and the optimizer = adam :

Epoch 1: val_accuracy improved from -inf to 0.56818, saving model to best_model.h5

Epoch 2: val_accuracy improved from 0.56818 to 0.65909, saving model to best_model.h5

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 3: val_accuracy improved from 0.65909 to 0.68182, saving model to best_model.h5

Epoch 4: val_accuracy improved from 0.68182 to 0.75000, saving model to best_model.h5

Epoch 5: val_accuracy improved from 0.75000 to 0.77273, saving model to best_model.h5

Epoch 6: val_accuracy did not improve from 0.77273

Epoch 7: val_accuracy improved from 0.77273 to 0.79545, saving model to best_model.h5

Epoch 8: val_accuracy did not improve from 0.79545

Epoch 9: val_accuracy did not improve from 0.79545

Epoch 10: val_accuracy did not improve from 0.79545

Epoch 11: val_accuracy did not improve from 0.79545

Epoch 12: val_accuracy did not improve from 0.79545

Epoch 13: val_accuracy did not improve from 0.79545

Epoch 14: val_accuracy did not improve from 0.79545

Epoch 15: val_accuracy improved from 0.79545 to 0.81818, saving model to best_model.h5

Epoch 16: val_accuracy did not improve from 0.81818

Epoch 17: val_accuracy improved from 0.81818 to 0.84091, saving model to best_model.h5

Epoch 18: val_accuracy did not improve from 0.84091

Epoch 19: val_accuracy did not improve from 0.84091

Epoch 20: val_accuracy did not improve from 0.84091

Epoch 21: val_accuracy did not improve from 0.84091

Epoch 22: val_accuracy improved from 0.84091 to 0.86364, saving model to best_model.h5

Epoch 23: val_accuracy did not improve from 0.86364

Epoch 24: val_accuracy improved from 0.86364 to 0.90909, saving model to best_model.h5

Epoch 25: val_accuracy did not improve from 0.90909

Epoch 26: val_accuracy did not improve from 0.90909

Epoch 27: val_accuracy did not improve from 0.90909

Epoch 28: val_accuracy did not improve from 0.90909

Epoch 29: val_accuracy did not improve from 0.90909

Epoch 30: val_accuracy did not improve from 0.90909

Epoch 31: val_accuracy did not improve from 0.90909

Epoch 32: val_accuracy did not improve from 0.90909

Epoch 33: val_accuracy did not improve from 0.90909

Epoch 34: val_accuracy did not improve from 0.90909

Epoch 35: val_accuracy did not improve from 0.90909

Epoch 36: val_accuracy did not improve from 0.90909

Epoch 37: val_accuracy did not improve from 0.90909

Epoch 38: val_accuracy did not improve from 0.90909

Epoch 39: val_accuracy did not improve from 0.90909

Epoch 40: val_accuracy did not improve from 0.90909

Epoch 41: val_accuracy did not improve from 0.90909

Epoch 42: val_accuracy did not improve from 0.90909

Epoch 43: val_accuracy did not improve from 0.90909

Epoch 44: val_accuracy did not improve from 0.90909

2/2 - 0s - loss: 0.4113 - accuracy: 0.9091 - 23ms/epoch - 12ms/step


#######################################################


the model mod6 use a learning rate = 3, l2 regularization = 0 and the optimizer = adam :

Epoch 1: val_accuracy improved from -inf to 0.25000, saving model to best_model.h5

Epoch 2: val_accuracy did not improve from 0.25000

Epoch 3: val_accuracy did not improve from 0.25000

Epoch 4: val_accuracy did not improve from 0.25000


classement=sorted(classement, reverse = True)
 print(str(i+1)+"-" + str(classement[i][1])+" with an accuracy of :"+ str(round(classement[i][0]*100,3)))

1-mod82 with an accuracy of :90.909
2-mod60 with an accuracy of :90.909
3-mod5 with an accuracy of :90.909
4-mod28 with an accuracy of :90.909
5-mod158 with an accuracy of :90.909
6-mod104 with an accuracy of :90.909
7-mod97 with an accuracy of :88.636
8-mod87 with an accuracy of :88.636
9-mod84 with an accuracy of :88.636
10-mod8 with an accuracy of :88.636
11-mod67 with an accuracy of :88.636
12-mod56 with an accuracy of :88.636
13-mod55 with an accuracy of :88.636
14-mod44 with an accuracy of :88.636
15-mod43 with an accuracy of :88.636
16-mod38 with an accuracy of :88.636
17-mod37 with an accuracy of :88.636
18-mod36 with an accuracy of :88.636
19-mod35 with an accuracy of :88.636
20-mod25 with an accuracy of :88.636
21-mod157 with an accuracy of :88.636
22-mod14 with an accuracy of :88.636
23-mod127 with an accuracy of :88.636
24-mod119 with an accuracy of :88.636
25-mod92 with an accuracy of :86.364
26-mod90 with an accuracy of :86.364
27-mod68 with an accuracy of :86.364
28-mod65 with an accuracy of :86.364
29-mod53 with an accuracy of :86.364
30-mod45 with an accuracy of :86.364
31-mod4 with an accuracy of :86.364
32-mod26 with an accuracy of :86.364
33-mod182 with an accuracy of :86.364
34-mod164 with an accuracy of :86.364
35-mod156 with an accuracy of :86.364
36-mod152 with an accuracy of :86.364
37-mod15 with an accuracy of :86.364
38-mod147 with an accuracy of :86.364
39-mod137 with an accuracy of :86.364
40-mod126 with an accuracy of :86.364
41-mod125 with an accuracy of :86.364
42-mod120 with an accuracy of :86.364
43-mod96 with an accuracy of :84.091
44-mod95 with an accuracy of :84.091
45-mod91 with an accuracy of :84.091
46-mod88 with an accuracy of :84.091
47-mod85 with an accuracy of :84.091
48-mod83 with an accuracy of :84.091
49-mod79 with an accuracy of :84.091
50-mod74 with an accuracy of :84.091
51-mod62 with an accuracy of :84.091
52-mod57 with an accuracy of :84.091
53-mod31 with an accuracy of :84.091
54-mod177 with an accuracy of :84.091
55-mod176 with an accuracy of :84.091
56-mod174 with an accuracy of :84.091
57-mod163 with an accuracy of :84.091
58-mod155 with an accuracy of :84.091
59-mod149 with an accuracy of :84.091
60-mod135 with an accuracy of :84.091
61-mod128 with an accuracy of :84.091
62-mod118 with an accuracy of :84.091
63-mod98 with an accuracy of :81.818
64-mod94 with an accuracy of :81.818
65-mod89 with an accuracy of :81.818
66-mod86 with an accuracy of :81.818
67-mod75 with an accuracy of :81.818
68-mod64 with an accuracy of :81.818
69-mod61 with an accuracy of :81.818
70-mod59 with an accuracy of :81.818
71-mod58 with an accuracy of :81.818
72-mod54 with an accuracy of :81.818
73-mod34 with an accuracy of :81.818
74-mod32 with an accuracy of :81.818
75-mod30 with an accuracy of :81.818
76-mod27 with an accuracy of :81.818
77-mod24 with an accuracy of :81.818
78-mod180 with an accuracy of :81.818
79-mod175 with an accuracy of :81.818
80-mod173 with an accuracy of :81.818
81-mod17 with an accuracy of :81.818
82-mod165 with an accuracy of :81.818
83-mod154 with an accuracy of :81.818
84-mod151 with an accuracy of :81.818
85-mod148 with an accuracy of :81.818
86-mod144 with an accuracy of :81.818
87-mod116 with an accuracy of :81.818
88-mod115 with an accuracy of :81.818
89-mod114 with an accuracy of :81.818
90-mod77 with an accuracy of :79.545
91-mod73 with an accuracy of :79.545
92-mod66 with an accuracy of :79.545
93-mod6 with an accuracy of :79.545
94-mod29 with an accuracy of :79.545
95-mod178 with an accuracy of :79.545
96-mod153 with an accuracy of :79.545
97-mod150 with an accuracy of :79.545
98-mod146 with an accuracy of :79.545
99-mod134 with an accuracy of :79.545
100-mod124 with an accuracy of :79.545
101-mod122 with an accuracy of :79.545
102-mod113 with an accuracy of :79.545
103-mod105 with an accuracy of :79.545
104-mod78 with an accuracy of :77.273
105-mod72 with an accuracy of :77.273
106-mod7 with an accuracy of :77.273
107-mod181 with an accuracy of :77.273
108-mod179 with an accuracy of :77.273
109-mod167 with an accuracy of :77.273
110-mod145 with an accuracy of :77.273
111-mod143 with an accuracy of :77.273
112-mod117 with an accuracy of :77.273
113-mod107 with an accuracy of :77.273
114-mod70 with an accuracy of :75.0
115-mod51 with an accuracy of :75.0
116-mod121 with an accuracy of :75.0
117-mod108 with an accuracy of :75.0
118-mod106 with an accuracy of :75.0
119-mod103 with an accuracy of :75.0
120-mod102 with an accuracy of :75.0
121-mod9 with an accuracy of :72.727
122-mod169 with an accuracy of :72.727
123-mod138 with an accuracy of :72.727
124-mod133 with an accuracy of :72.727
125-mod47 with an accuracy of :70.455
126-mod130 with an accuracy of :70.455
127-mod10 with an accuracy of :70.455
128-mod48 with an accuracy of :68.182
129-mod161 with an accuracy of :68.182
130-mod16 with an accuracy of :68.182
131-mod159 with an accuracy of :68.182
132-mod123 with an accuracy of :68.182
133-mod11 with an accuracy of :68.182
134-mod69 with an accuracy of :65.909
135-mod131 with an accuracy of :65.909
136-mod129 with an accuracy of :65.909
137-mod101 with an accuracy of :65.909
138-mod23 with an accuracy of :63.636
139-mod112 with an accuracy of :63.636
140-mod93 with an accuracy of :61.364
141-mod49 with an accuracy of :61.364
142-mod172 with an accuracy of :61.364
143-mod76 with an accuracy of :59.091
144-mod168 with an accuracy of :59.091
145-mod100 with an accuracy of :59.091
146-mod20 with an accuracy of :56.818
147-mod162 with an accuracy of :56.818
148-mod80 with an accuracy of :54.545
149-mod63 with an accuracy of :54.545
150-mod46 with an accuracy of :54.545
151-mod33 with an accuracy of :54.545
152-mod3 with an accuracy of :54.545
153-mod22 with an accuracy of :54.545
154-mod19 with an accuracy of :54.545
155-mod41 with an accuracy of :52.273
156-mod141 with an accuracy of :52.273
157-mod52 with an accuracy of :50.0
158-mod18 with an accuracy of :50.0
159-mod166 with an accuracy of :50.0
160-mod139 with an accuracy of :50.0
161-mod13 with an accuracy of :50.0
162-mod81 with an accuracy of :47.727
163-mod12 with an accuracy of :47.727
164-mod111 with an accuracy of :45.455
165-mod99 with an accuracy of :43.182
166-mod40 with an accuracy of :40.909
167-mod136 with an accuracy of :40.909
168-mod109 with an accuracy of :40.909
169-mod160 with an accuracy of :38.636
170-mod21 with an accuracy of :36.364
171-mod171 with an accuracy of :36.364
172-mod142 with an accuracy of :36.364
173-mod140 with an accuracy of :36.364
174-mod110 with an accuracy of :34.091
175-mod50 with an accuracy of :31.818
176-mod170 with an accuracy of :31.818
177-mod132 with an accuracy of :29.545
178-mod71 with an accuracy of :27.273
179-mod42 with an accuracy of :22.727
180-mod39 with an accuracy of :18.182


model = keras.Sequential([
    keras.layers.Dense(13, input_dim=13, activation='relu', kernel_initializer='he_uniform',kernel_regularizer=keras.regularizers.l2(3)),
    #13 neurons because it's the number of columns/inputs
    keras.layers.Dense(2, activation='sigmoid')
    #2 neurons because we want a classification with two labels
])
es = EarlyStopping(monitor='val_accuracy', mode='auto', verbose=0,patience=20)
mc = ModelCheckpoint('best_model.h5', monitor='val_accuracy', mode='max', verbose=1, save_best_only=True)
opt=Adagrad(learning_rate= 1)
model.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])
history = model.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=1800, verbose=0, callbacks=[es, mc])

Epoch 1: val_accuracy improved from -inf to 0.68182, saving model to best_model.h5

Epoch 2: val_accuracy did not improve from 0.68182

Epoch 3: val_accuracy did not improve from 0.68182

Epoch 4: val_accuracy improved from 0.68182 to 0.79545, saving model to best_model.h5

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(

Epoch 5: val_accuracy did not improve from 0.79545

Epoch 6: val_accuracy did not improve from 0.79545

Epoch 7: val_accuracy did not improve from 0.79545

Epoch 8: val_accuracy did not improve from 0.79545

Epoch 9: val_accuracy did not improve from 0.79545

Epoch 10: val_accuracy improved from 0.79545 to 0.88636, saving model to best_model.h5

Epoch 11: val_accuracy did not improve from 0.88636

Epoch 12: val_accuracy did not improve from 0.88636

Epoch 13: val_accuracy did not improve from 0.88636

Epoch 14: val_accuracy did not improve from 0.88636

Epoch 15: val_accuracy did not improve from 0.88636

Epoch 16: val_accuracy did not improve from 0.88636

Epoch 17: val_accuracy did not improve from 0.88636

Epoch 18: val_accuracy did not improve from 0.88636

Epoch 19: val_accuracy did not improve from 0.88636

Epoch 20: val_accuracy did not improve from 0.88636

Epoch 21: val_accuracy did not improve from 0.88636

Epoch 22: val_accuracy did not improve from 0.88636

Epoch 23: val_accuracy did not improve from 0.88636

Epoch 24: val_accuracy improved from 0.88636 to 0.90909, saving model to best_model.h5

Epoch 25: val_accuracy did not improve from 0.90909

Epoch 26: val_accuracy did not improve from 0.90909

Epoch 27: val_accuracy did not improve from 0.90909

Epoch 28: val_accuracy did not improve from 0.90909

Epoch 29: val_accuracy did not improve from 0.90909

Epoch 30: val_accuracy did not improve from 0.90909

Epoch 31: val_accuracy did not improve from 0.90909

Epoch 32: val_accuracy did not improve from 0.90909

Epoch 33: val_accuracy did not improve from 0.90909

Epoch 34: val_accuracy did not improve from 0.90909

Epoch 35: val_accuracy did not improve from 0.90909

Epoch 36: val_accuracy did not improve from 0.90909

Epoch 37: val_accuracy did not improve from 0.90909

Epoch 38: val_accuracy did not improve from 0.90909

Epoch 39: val_accuracy did not improve from 0.90909

Epoch 40: val_accuracy did not improve from 0.90909

Epoch 41: val_accuracy did not improve from 0.90909

Epoch 42: val_accuracy did not improve from 0.90909

Epoch 43: val_accuracy did not improve from 0.90909

Epoch 44: val_accuracy did not improve from 0.90909


X = small_validation.drop(columns="disease")
y= small_validation["disease"]
X-= mean
X /= std
small_predictions = model.predict(X)
result =[]
for i in range(len(small_predictions)):
  if small_predictions[i][0]>small_predictions[i][1]:
    result.append(1)#the real class
  else :
    result.append(2)#the real class
score = accuracy_score(small_validation["disease"], result)
print("accuracy : "+str(score))

2/2 [==============================] - 0s 8ms/step
accuracy : 0.7777777777777778


#writing results for exam evaluation
r= open("results_small_dataset.txt", "w+")
r.write(str(result))
r.close()

	0	1	2	3	4	5	6	7	8	9	10	11	12	13
0	50.0	0.0	2.0	120.0	244.0	0.0	0.0	162.0	0.0	1.1	1.0	0.0	3.0	1
1	56.0	1.0	4.0	130.0	283.0	1.0	2.0	103.0	1.0	1.6	3.0	0.0	7.0	2
2	44.0	1.0	3.0	120.0	226.0	0.0	0.0	169.0	0.0	0.0	1.0	0.0	3.0	1
3	55.0	1.0	4.0	160.0	289.0	0.0	2.0	145.0	1.0	0.8	2.0	1.0	7.0	2
4	54.0	1.0	2.0	108.0	309.0	0.0	0.0	156.0	0.0	0.0	1.0	0.0	7.0	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
211	44.0	0.0	3.0	108.0	141.0	0.0	0.0	175.0	0.0	0.6	2.0	0.0	3.0	1
212	50.0	1.0	3.0	129.0	196.0	0.0	0.0	163.0	0.0	0.0	1.0	0.0	3.0	1
213	63.0	0.0	4.0	150.0	407.0	0.0	2.0	154.0	0.0	4.0	2.0	3.0	7.0	2
214	50.0	1.0	4.0	150.0	243.0	0.0	2.0	128.0	0.0	2.6	2.0	0.0	7.0	2
215	45.0	0.0	4.0	138.0	236.0	0.0	2.0	152.0	1.0	0.2	2.0	0.0	3.0	1

	age	sex	chest_pain	blood_pressure	cholestoral	blood_sugar	electrocardiographic	max_heart_rate	exercise	oldpeak	peak_exer	nb_vessel	defect	disease
0	50.0	0.0	2.0	120.0	244.0	0.0	0.0	162.0	0.0	1.1	1.0	0.0	3.0	1
1	56.0	1.0	4.0	130.0	283.0	1.0	2.0	103.0	1.0	1.6	3.0	0.0	7.0	2
2	44.0	1.0	3.0	120.0	226.0	0.0	0.0	169.0	0.0	0.0	1.0	0.0	3.0	1
3	55.0	1.0	4.0	160.0	289.0	0.0	2.0	145.0	1.0	0.8	2.0	1.0	7.0	2
4	54.0	1.0	2.0	108.0	309.0	0.0	0.0	156.0	0.0	0.0	1.0	0.0	7.0	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
211	44.0	0.0	3.0	108.0	141.0	0.0	0.0	175.0	0.0	0.6	2.0	0.0	3.0	1
212	50.0	1.0	3.0	129.0	196.0	0.0	0.0	163.0	0.0	0.0	1.0	0.0	3.0	1
213	63.0	0.0	4.0	150.0	407.0	0.0	2.0	154.0	0.0	4.0	2.0	3.0	7.0	2
214	50.0	1.0	4.0	150.0	243.0	0.0	2.0	128.0	0.0	2.6	2.0	0.0	7.0	2
215	45.0	0.0	4.0	138.0	236.0	0.0	2.0	152.0	1.0	0.2	2.0	0.0	3.0	1

	Unnamed: 0	age	sex	chest_pain	blood_pressure	cholestoral	blood_sugar	electrocardiographic	max_heart_rate	exercise	oldpeak	peak_exer	nb_vessel	defect	disease
0	0	50.0	0.0	2.0	120.0	244.0	0.0	0.0	162.0	0.0	1.1	1.0	0.0	3.0	1
1	1	56.0	1.0	4.0	130.0	283.0	1.0	2.0	103.0	1.0	1.6	3.0	0.0	7.0	2
2	2	44.0	1.0	3.0	120.0	226.0	0.0	0.0	169.0	0.0	0.0	1.0	0.0	3.0	1
3	3	55.0	1.0	4.0	160.0	289.0	0.0	2.0	145.0	1.0	0.8	2.0	1.0	7.0	2
4	4	54.0	1.0	2.0	108.0	309.0	0.0	0.0	156.0	0.0	0.0	1.0	0.0	7.0	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
7555	211	44.0	0.0	3.0	108.0	141.0	0.0	0.0	175.0	0.0	0.6	2.0	0.0	3.0	1
7556	212	50.0	1.0	3.0	129.0	196.0	0.0	0.0	163.0	0.0	0.0	1.0	0.0	3.0	1
7557	213	63.0	0.0	4.0	150.0	407.0	0.0	2.0	154.0	0.0	4.0	2.0	3.0	7.0	2
7558	214	50.0	1.0	4.0	150.0	243.0	0.0	2.0	128.0	0.0	2.6	2.0	0.0	7.0	2
7559	215	45.0	0.0	4.0	138.0	236.0	0.0	2.0	152.0	1.0	0.2	2.0	0.0	3.0	1

	0	1	2	3	4	5	6	7	8	9	10	11	12	13
0	47.0	1.0	4.0	112.0	204.0	0.0	0.0	143.0	0.0	0.1	1.0	0.0	3.0	1
1	62.0	0.0	4.0	138.0	294.0	1.0	0.0	106.0	0.0	1.9	2.0	3.0	3.0	2
2	55.0	1.0	4.0	160.0	289.0	0.0	2.0	145.0	1.0	0.8	2.0	1.0	7.0	2
3	71.0	0.0	4.0	112.0	149.0	0.0	0.0	125.0	0.0	1.6	2.0	0.0	3.0	1
4	67.0	0.0	3.0	115.0	564.0	0.0	2.0	160.0	0.0	1.6	2.0	0.0	7.0	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
211	56.0	1.0	1.0	120.0	193.0	0.0	2.0	162.0	0.0	1.9	2.0	0.0	7.0	1
212	49.0	0.0	2.0	134.0	271.0	0.0	0.0	162.0	0.0	0.0	2.0	0.0	3.0	1
213	54.0	0.0	3.0	110.0	214.0	0.0	0.0	158.0	0.0	1.6	2.0	0.0	3.0	1
214	41.0	0.0	2.0	130.0	204.0	0.0	2.0	172.0	0.0	1.4	1.0	0.0	3.0	1
215	66.0	0.0	4.0	178.0	228.0	1.0	0.0	165.0	1.0	1.0	2.0	2.0	7.0	2

	age	sex	chest_pain	blood_pressure	cholestoral	blood_sugar	electrocardiographic	max_heart_rate	exercise	oldpeak	peak_exer	nb_vessel	defect	disease
0	47.0	1.0	4.0	112.0	204.0	0.0	0.0	143.0	0.0	0.1	1.0	0.0	3.0	1
1	62.0	0.0	4.0	138.0	294.0	1.0	0.0	106.0	0.0	1.9	2.0	3.0	3.0	2
2	55.0	1.0	4.0	160.0	289.0	0.0	2.0	145.0	1.0	0.8	2.0	1.0	7.0	2
3	71.0	0.0	4.0	112.0	149.0	0.0	0.0	125.0	0.0	1.6	2.0	0.0	3.0	1
4	67.0	0.0	3.0	115.0	564.0	0.0	2.0	160.0	0.0	1.6	2.0	0.0	7.0	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
211	56.0	1.0	1.0	120.0	193.0	0.0	2.0	162.0	0.0	1.9	2.0	0.0	7.0	1
212	49.0	0.0	2.0	134.0	271.0	0.0	0.0	162.0	0.0	0.0	2.0	0.0	3.0	1
213	54.0	0.0	3.0	110.0	214.0	0.0	0.0	158.0	0.0	1.6	2.0	0.0	3.0	1
214	41.0	0.0	2.0	130.0	204.0	0.0	2.0	172.0	0.0	1.4	1.0	0.0	3.0	1
215	66.0	0.0	4.0	178.0	228.0	1.0	0.0	165.0	1.0	1.0	2.0	2.0	7.0	2

Deep learning project :¶

Libraries :¶

Importation of the dataset :¶

Large dataset:¶

Training sets :¶

Validation set :¶

Small dataset¶

Training dataset :¶

Validation :¶

first exploration of the dataset:¶

Data cleaning for deep learning modelisation:¶

Large dataset :¶

Small dataset :¶

Modelizations and tests :¶

Useful functions :¶

Modelizations with the large dataset :¶

Validation and final prediction:¶

Modelizations with the small dataset :¶

Validation and final prediction:¶

	age	sex	chest_pain	blood_pressure	cholestoral	blood_sugar	electrocardiographic	max_heart_rate	exercise	oldpeak	peak_exer	nb_vessel	defect	disease
0	67.0	0.0	4.0	106.0	223.0	0.0	0.0	142.0	0.0	0.3	1.0	2.0	3.0	1
1	52.0	1.0	2.0	134.0	201.0	0.0	0.0	158.0	0.0	0.8	1.0	1.0	3.0	1
2	50.0	0.0	2.0	120.0	244.0	0.0	0.0	162.0	0.0	1.1	1.0	0.0	3.0	1
3	57.0	1.0	4.0	165.0	289.0	1.0	2.0	124.0	0.0	1.0	2.0	3.0	7.0	2
4	48.0	1.0	3.0	124.0	255.0	1.0	0.0	175.0	0.0	0.0	1.0	2.0	3.0	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
211	42.0	1.0	3.0	130.0	180.0	0.0	0.0	150.0	0.0	0.0	1.0	0.0	3.0	1
212	57.0	0.0	4.0	120.0	354.0	0.0	0.0	163.0	1.0	0.6	1.0	0.0	3.0	1
213	65.0	0.0	4.0	150.0	225.0	0.0	2.0	114.0	0.0	1.0	2.0	3.0	7.0	2
214	38.0	1.0	1.0	120.0	231.0	0.0	0.0	182.0	1.0	3.8	2.0	0.0	7.0	2
215	47.0	1.0	4.0	112.0	204.0	0.0	0.0	143.0	0.0	0.1	1.0	0.0	3.0	1

	0	1	2	3	4	5	6	7	8	9	10	11	12	13
0	63.0	0.0	3.0	135.0	252.0	0.0	2.0	172.0	0.0	0.0	1.0	0.0	3.0	1
1	51.0	1.0	3.0	94.0	227.0	0.0	0.0	154.0	1.0	0.0	1.0	1.0	7.0	1
2	54.0	1.0	3.0	120.0	258.0	0.0	2.0	147.0	0.0	0.4	2.0	0.0	7.0	1
3	44.0	1.0	2.0	120.0	220.0	0.0	0.0	170.0	0.0	0.0	1.0	0.0	3.0	1
4	54.0	1.0	4.0	110.0	239.0	0.0	0.0	126.0	1.0	2.8	2.0	1.0	7.0	2
5	65.0	1.0	4.0	135.0	254.0	0.0	2.0	127.0	0.0	2.8	2.0	1.0	7.0	2
6	57.0	1.0	3.0	150.0	168.0	0.0	0.0	174.0	0.0	1.6	1.0	0.0	3.0	1
7	63.0	1.0	4.0	130.0	330.0	1.0	2.0	132.0	1.0	1.8	1.0	3.0	7.0	2
8	35.0	0.0	4.0	138.0	183.0	0.0	0.0	182.0	0.0	1.4	1.0	0.0	3.0	1
9	41.0	1.0	2.0	135.0	203.0	0.0	0.0	132.0	0.0	0.0	2.0	0.0	6.0	1
10	62.0	0.0	3.0	130.0	263.0	0.0	0.0	97.0	0.0	1.2	2.0	1.0	7.0	2
11	43.0	0.0	4.0	132.0	341.0	1.0	2.0	136.0	1.0	3.0	2.0	0.0	7.0	2
12	58.0	0.0	1.0	150.0	283.0	1.0	2.0	162.0	0.0	1.0	1.0	0.0	3.0	1
13	52.0	1.0	1.0	118.0	186.0	0.0	2.0	190.0	0.0	0.0	2.0	0.0	6.0	1
14	61.0	0.0	4.0	145.0	307.0	0.0	2.0	146.0	1.0	1.0	2.0	0.0	7.0	2
15	39.0	1.0	4.0	118.0	219.0	0.0	0.0	140.0	0.0	1.2	2.0	0.0	7.0	2
16	45.0	1.0	4.0	115.0	260.0	0.0	2.0	185.0	0.0	0.0	1.0	0.0	3.0	1
17	52.0	1.0	4.0	128.0	255.0	0.0	0.0	161.0	1.0	0.0	1.0	1.0	7.0	2
18	62.0	1.0	3.0	130.0	231.0	0.0	0.0	146.0	0.0	1.8	2.0	3.0	7.0	1
19	62.0	0.0	4.0	160.0	164.0	0.0	2.0	145.0	0.0	6.2	3.0	3.0	7.0	2
20	53.0	0.0	4.0	138.0	234.0	0.0	2.0	160.0	0.0	0.0	1.0	0.0	3.0	1
21	43.0	1.0	4.0	120.0	177.0	0.0	2.0	120.0	1.0	2.5	2.0	0.0	7.0	2
22	47.0	1.0	3.0	138.0	257.0	0.0	2.0	156.0	0.0	0.0	1.0	0.0	3.0	1
23	52.0	1.0	2.0	120.0	325.0	0.0	0.0	172.0	0.0	0.2	1.0	0.0	3.0	1
24	68.0	1.0	3.0	180.0	274.0	1.0	2.0	150.0	1.0	1.6	2.0	0.0	7.0	2
25	39.0	1.0	3.0	140.0	321.0	0.0	2.0	182.0	0.0	0.0	1.0	0.0	3.0	1
26	53.0	0.0	4.0	130.0	264.0	0.0	2.0	143.0	0.0	0.4	2.0	0.0	3.0	1
27	62.0	0.0	4.0	140.0	268.0	0.0	2.0	160.0	0.0	3.6	3.0	2.0	3.0	2
28	51.0	0.0	3.0	140.0	308.0	0.0	2.0	142.0	0.0	1.5	1.0	1.0	3.0	1
29	60.0	1.0	4.0	130.0	253.0	0.0	0.0	144.0	1.0	1.4	1.0	1.0	7.0	2
30	65.0	1.0	4.0	110.0	248.0	0.0	2.0	158.0	0.0	0.6	1.0	2.0	6.0	2
31	65.0	0.0	3.0	155.0	269.0	0.0	0.0	148.0	0.0	0.8	1.0	0.0	3.0	1
32	60.0	1.0	3.0	140.0	185.0	0.0	2.0	155.0	0.0	3.0	2.0	0.0	3.0	2
33	60.0	1.0	4.0	145.0	282.0	0.0	2.0	142.0	1.0	2.8	2.0	2.0	7.0	2
34	54.0	1.0	4.0	120.0	188.0	0.0	0.0	113.0	0.0	1.4	2.0	1.0	7.0	2
35	44.0	1.0	2.0	130.0	219.0	0.0	2.0	188.0	0.0	0.0	1.0	0.0	3.0	1
36	44.0	1.0	4.0	112.0	290.0	0.0	2.0	153.0	0.0	0.0	1.0	1.0	3.0	2
37	51.0	1.0	3.0	110.0	175.0	0.0	0.0	123.0	0.0	0.6	1.0	0.0	3.0	1
38	59.0	1.0	3.0	150.0	212.0	1.0	0.0	157.0	0.0	1.6	1.0	0.0	3.0	1
39	71.0	0.0	2.0	160.0	302.0	0.0	0.0	162.0	0.0	0.4	1.0	2.0	3.0	1
40	61.0	1.0	3.0	150.0	243.0	1.0	0.0	137.0	1.0	1.0	2.0	0.0	3.0	1
41	55.0	1.0	4.0	132.0	353.0	0.0	0.0	132.0	1.0	1.2	2.0	1.0	7.0	2
42	64.0	1.0	3.0	140.0	335.0	0.0	0.0	158.0	0.0	0.0	1.0	0.0	3.0	2
43	43.0	1.0	4.0	150.0	247.0	0.0	0.0	171.0	0.0	1.5	1.0	0.0	3.0	1
44	58.0	0.0	3.0	120.0	340.0	0.0	0.0	172.0	0.0	0.0	1.0	0.0	3.0	1
45	60.0	1.0	4.0	130.0	206.0	0.0	2.0	132.0	1.0	2.4	2.0	2.0	7.0	2
46	58.0	1.0	2.0	120.0	284.0	0.0	2.0	160.0	0.0	1.8	2.0	0.0	3.0	2
47	49.0	1.0	2.0	130.0	266.0	0.0	0.0	171.0	0.0	0.6	1.0	0.0	3.0	1
48	48.0	1.0	2.0	110.0	229.0	0.0	0.0	168.0	0.0	1.0	3.0	0.0	7.0	2
49	52.0	1.0	3.0	172.0	199.0	1.0	0.0	162.0	0.0	0.5	1.0	0.0	7.0	1
50	44.0	1.0	2.0	120.0	263.0	0.0	0.0	173.0	0.0	0.0	1.0	0.0	7.0	1
51	56.0	0.0	2.0	140.0	294.0	0.0	2.0	153.0	0.0	1.3	2.0	0.0	3.0	1
52	57.0	1.0	4.0	140.0	192.0	0.0	0.0	148.0	0.0	0.4	2.0	0.0	6.0	1
53	67.0	1.0	4.0	160.0	286.0	0.0	2.0	108.0	1.0	1.5	2.0	3.0	3.0	2

	age	sex	chest_pain	blood_pressure	cholestoral	blood_sugar	electrocardiographic	max_heart_rate	exercise	oldpeak	peak_exer	nb_vessel	defect	disease
count	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000	7560.000000
mean	54.449074	0.666667	3.171296	130.777778	248.967593	0.152778	1.027778	149.148148	0.347222	1.026389	1.597222	0.689815	4.657407	1.444444
std	9.249095	0.471436	0.968697	18.083740	52.246423	0.359797	0.995038	23.757785	0.476119	1.129028	0.616025	0.943406	1.935003	0.496937
min	29.000000	0.000000	1.000000	94.000000	126.000000	0.000000	0.000000	71.000000	0.000000	0.000000	1.000000	0.000000	3.000000	1.000000
25%	47.750000	0.000000	3.000000	120.000000	212.750000	0.000000	0.000000	131.750000	0.000000	0.000000	1.000000	0.000000	3.000000	1.000000
50%	55.000000	1.000000	3.000000	130.000000	243.000000	0.000000	2.000000	153.500000	0.000000	0.800000	2.000000	0.000000	3.000000	1.000000
75%	61.000000	1.000000	4.000000	140.000000	276.250000	0.000000	2.000000	166.000000	1.000000	1.800000	2.000000	1.000000	7.000000	2.000000
max	77.000000	1.000000	4.000000	200.000000	564.000000	1.000000	2.000000	202.000000	1.000000	5.600000	3.000000	3.000000	7.000000	2.000000

	age	sex	chest_pain	blood_pressure	cholestoral	blood_sugar	electrocardiographic	max_heart_rate	exercise	oldpeak	peak_exer	nb_vessel	defect
14	59.0	1.0	2.0	140.0	221.0	0.0	0.0	164.0	1.0	0.0	1.0	0.0	3.0
137	65.0	1.0	4.0	120.0	177.0	0.0	0.0	140.0	0.0	0.4	1.0	0.0	7.0
164	69.0	1.0	3.0	140.0	254.0	0.0	2.0	146.0	0.0	2.0	2.0	3.0	7.0
97	57.0	1.0	3.0	150.0	126.0	1.0	0.0	173.0	0.0	0.2	1.0	1.0	7.0
84	63.0	0.0	4.0	108.0	269.0	0.0	0.0	169.0	1.0	1.8	2.0	2.0	3.0